fix(publisher): skip when failed to retrieve container

There is a race condition where a container starts via `docker run`,
but then immediately fails because an incorrect command was specified
or some other error occurred that caused the container to fail over. In
this case, we should just log and fail over, rather than killing
publisher.

Author: Matthew Fisher

publisher/publisher.go

@@ -60,7 +60,8 @@ func listenContainers(client *docker.Client, etcdClient *etcd.Client, ttl time.D
 			if event.Status == "start" {
 				container, err := getContainer(client, event.ID)
 				if err != nil {
-					log.Fatal(err)
+					log.Println(err)
+					continue
 				}
 				publishContainer(etcdClient, container, ttl)
 			}